Vision Box DAYTONA

  • Embedded Box-PC supporting Nvidia Jetson TX2
  • Suitable for deep learning inference and other applications requiring a GPU accelerator
  • Supports vision relevant interfaces
  • Compact, low power consumption, long-term availability

Several applications require an accelerator which understands CUDA based software. For the industry, short-term available high power-consumption GPU cards are not the optimal solution. NVidia offers with the Jetson TX2 the combination of a multi-core ARM-based CPU and 256 GPU shader able to run your CUDA code.

But a vision application does not need just a CPU and GPU – it needs vision relevant interfaces. The GPGPU computer Vision Box DAYTONA serves all the requirements: cameras can be connected with a single cable able to transport the data, power and especially the trigger signal. IO and an encoder interface support sensors and actors. And finally, for easy access from outside, a modem can be connected directly to your support team.

All the functions are covered in an industrial, fanless, low power consumption housing.

The fanless vision computer Vision Box DAYTONA – image processing hardware for the following applications:

  • Deep Learning Inference Programs. While the learning phase of neural networks runs offline in the cloud or on a specialized computer, the executable program of deep learning, so-called inference programs runs on the Vision Box DAYTONA and makes it a powerful inference computer.
  • Hyperspectral Imaging: This is a new dimension of big data coming from cameras able to „see“ what the human doesn’t see. Up to 1 700 nm wavelength, SWIR based cameras create data to analyze e.g. material. The related algorithms need a CUDA based accelerator as it is available in the Vision Box DAYTONA.
  • Lightfield cameras: They are used in 3D applications and create many images from an object. To compute the amount of data, a GPU accelerator is welcome, too.

Suitable for AI-based image processing and other applications requiring an GPU accelerator
Regarding the application development, first, you have a multi-core Linux computer on your desk. Under the Linux OS, you develop the application with the freedom to use CUDA operators running on the GPU.

Just to underline the differences: With the Linux OS, IMAGO provides Vision Boxes with x86 (i-Core) CPU inside, alternatively an 8-Core ARM Cortex A72 CPU. The Vision Box DAYTONA instead uses a Quad-Core ARM A57 together with a 256 shader GPU.

Connect your camera, sensors, PLC and ethernet interface. IO are managed by our real-time communication controller – covered in an SDK for Linux OS. Start to develop your application.

Supports vision relevant interfaces
GigE Power and Trigger over Ethernet: A single ethernet cable supporting all camera functions as power supply, image data transmission, setup of parameter and especially IMAGO’s trigger over ethernet which is supported by a number of camera makers.

Real-Time IO: Many applications do not need a hard real-time OS – but for peripherals, it is very important to receive input data, to create new output data via a logic and to serve the process timing as many machines are working very fast. IMAGO’s real-time communication controller is a key feature to guarantee your process timing and allows real-time GPU processing.

Incremental Encoder Input: You need the speed of the conveyer belt? Not only for line scan applications but also together with area scan cameras, the information of the speed can become important. The Encoder Input is connected to the real-time communication controller with several logical functions to ensure a reliable application.

HDMI output: Connect directly a display to manage your software.

Compact, low power consumption, long-term availability
The image processing computer Vision Box DAYTONA is only 163mm x 163mm x 48mm in size. The compact housing just consumes less than 25W. Computing power is combined with low power consumption. As with many IMAGO products, also the Deep Learning Box is long-term available.

Food & Beverage
For example, in the food and beverage industry, more complex (deep learning) or big data applications (hyperspectral) become of interest. Integration into a stainless steel housing is easy thanks to the compact form factor and low-power consumption.
Further sample applications

For example, hyperspectral analysis becomes important for pharmaceutical products. Once the applications run under the Linux OS, all validation processes become easier as the OS can be freezed? for a long time. The Vision Box DAYTONA covers your application, nothing more.
Further sample applications

In the production process, deep learning becomes more important. As an example, in the die casting industry, deep learning-based software is already used. To support self-sufficient applications, the HDMI output for the monitor is welcome.
Further sample applications

Packaging & Logistics
For example, a computer installed on a forklift? Why not using the Vision Box DAYTONA, connect cameras and run more intelligent programs able to extract more data to control the forklift.

For example, with the GPU accelerator, e.g. code reading applications can be improved as in the logistics industry more and more codes are on packages.

For example, packaging machines themselves welcome the real-time behavior of the IO as they run very fast.
Further sample applications

Developers of machine vision applications using CUDA based software. Software engineers who are familiar with the Linux OS and Nvidia based processors.

Oliver Barz IMAGO Sales ManagerOliver Barz
Sales Manager

Telephone: +49 6031-684 26 11

Oliver Barz IMAGO Sales ManagerChristoph Siemon
Sales Manager

Telephone: +49 6031-684 26 11